2017-06-05 01:25:37

by Jia-Ju Bai

Subject: [PATCH resend] rt5208: Fix a sleep-in-atomic bug in xd_copy_page

Last patch lacks code explanation, and it is included in this patch.

The driver may sleep under a spin lock, and the function call path is:
rtsx_exclusive_enter_ss (acquire the lock by spin_lock)
rtsx_enter_ss
rtsx_power_off_card
xd_cleanup_work
xd_delay_write
xd_finish_write
xd_copy_page
wait_timeout
schedule_timeout --> may sleep

To fix it, "wait_timeout" is replaced with mdelay in xd_copy_page.

Signed-off-by: Jia-Ju Bai <[email protected]>
---
drivers/staging/rts5208/xd.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/rts5208/xd.c b/drivers/staging/rts5208/xd.c
index 85aba05..74d36f9 100644
--- a/drivers/staging/rts5208/xd.c
+++ b/drivers/staging/rts5208/xd.c
@@ -1268,7 +1268,7 @@ static int xd_copy_page(struct rtsx_chip *chip, u32 old_blk, u32 new_blk,
reg = 0;
rtsx_read_register(chip, XD_CTL, &reg);
if (reg & (XD_ECC1_ERROR | XD_ECC2_ERROR)) {
- wait_timeout(100);
+ mdelay(100);

if (detect_card_cd(chip,
XD_CARD) != STATUS_SUCCESS) {
--
1.7.9.5

2017-06-05 06:35:40

by Greg Kroah-Hartman

Subject: Re: [PATCH resend] rt5208: Fix a sleep-in-atomic bug in xd_copy_page

On Mon, Jun 05, 2017 at 09:26:35AM +0800, Jia-Ju Bai wrote:
> Last patch lacks code explanation, and it is included in this patch.

That should go below the --- line, as a 'v2' patch. Please fix it up so
I don't have to hand-edit this and resend.

Please read Documentation/SubmittingPatches for how to properly version
your patches. And also, please put the subsystem in your subject line,
otherwise it's hard for me to find them at times. Look at how other
patches subject lines are for these same drivers (i.e. you need
'staging' in it...)

thanks,

greg k-h